Among children less than 12 years of age residing in Dallas County, Texas, and in the state of Minnesota we conducted prospective, active surveillance of invasive Haemophilus influenzae disease. During 18 months, 616 cases were identified, of which 600 were caused by type b organisms. The annual incidence of disease was significantly greater in Dallas than in Minnesota (109 v 68/100,000 children younger than 5 years of age, P less than .001) and was greater in Dallas, even when rates for white children in the two regions were compared (P less than .001). Other regional differences were observed. In Dallas, a larger proportion of cases were in children attending day-care centers (27% compared with 12% in Minnesota, P less than .001) and more patients attended day care for greater than 40 h/wk (56% compared with 30% in Minnesota, P less than .001). Outer membrane protein subtyping of isolates revealed that in Dallas 6U isolates were associated significantly with cases in black children who attended day care. In Minnesota, but not in Dallas, isolates with subtype 1H were associated significantly with cases in children in day care. These data indicate that there are regional differences in the epidemiology of type b Haemophilus disease that may relate to differences in strains, day-care practices, or other unknown cultural or environmental factors. Finally, because only 15% of systemic Haemophilus disease in these regions occurred in children in the age groups recommended for vaccination (24 to 59 months), the new Haemophilus type b polysaccharide vaccine is expected to have a limited impact on the overall incidence of disease.